HIV rebound after interruption of anti-retroviral therapy (ART) is generally thought to reflect the size of the persistent reservoir found in long-lived memory CD4+ T cells, with a smaller reservoir reflected in delayed rebound. However, tissue reservoirs under ART suppression have been poorly characterized, so we do not know if characteristics of those reservoirs have a dominant impact on rebound kinetics. The contribution of host immune responses to rebound kinetics is also poorly understood, especially since most available data are focused on IFN-? producing T cell responses that are not related to control of viral load in acute or chronic infection. It would be surprising if the indiscriminate total of such responses had an important impact on rebound. Project 2 will leverage unique advantages of the rhesus macaque model to understand how tissue reservoirs and host immunity impact rebound. The animal model will provide frequent longitudinal samples of tissue-resident cells and recrudescent virus throughout suppressive ART and during rebound, which will be characterized by Dr. Mario Stevenson under the auspices of Project 3. In addition, we have the unique capability in rhesus macaques to induce both Mamu-E-restricted CD8+ T cells and memory NK cells using viral IL-10-deleted RhCMV/SIV vaccines. In contrast to conventional class Ia-restricted CD8+ T cells expressing IFN-?,1,2 Mamu-E-restricted CD8+ T cells likely contribute to suppression and even elimination of virus soon after infection, when SIV is replicating quickly in T cells, many of which are confined to gut tissue. We propose that the situation in a human patient after ART withdrawal is similar to that early stage of infection, in that HIV rebound presumably begins from a small number of cells reactivating virus in restricted anatomic sites. Thus, if HLA-E-restricted responses are present in some human patients (Project 1) or can be induced in vaccinated humans or macaques, these individuals might exhibit delayed rebound. We hypothesize that the unique immune responses elicited by RhCMV/SIV therapeutic vaccination under ART, but not conventional CTL elicited by Ad/SIV vaccine when used alone, alter both tissue viral reservoirs and SIV rebound kinetics. Our specific aims are: 1. Test if therapeutic RhCMV/SIV vaccination is superior to Ad/SIV vaccination for depletion of the Tfh-resident SIV reservoir. 2. Test if RhCMV/SIV vaccination slows rebound kinetics. 3. Determine if HLA/Mamu-E-restricted T cell responses are induced in unvaccinated individuals by CMV co-infection and are a biomarker for delayed rebound.